Table 6 Classification metrics comparison of the relevant classes between the US8K and US8K_AV datasets using fold 1 as validation set.
From: A dataset for environmental sound recognition in embedded systems for autonomous vehicles
US8K | US8K_AV | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
Model | Class | Prec. | Recall | F1-score | Prec. | Recall | F1-score | Suppt. | Imp. | Wtd. |
SVM | car_horn | 64% | 88% | 74% | 81% | 96% | 88% | 252 | 19% | 9.58% |
children_playing | 58% | 82% | 68% | 71% | 81% | 76% | 700 | 12% | ||
dog_bark | 76% | 78% | 77% | 85% | 78% | 81% | 700 | 5% | ||
siren | 82% | 65% | 73% | 85% | 75% | 79% | 602 | 8% | ||
LR | car_horn | 67% | 85% | 75% | 84% | 86% | 82% | 252 | 9% | 6.36% |
children_playing | 63% | 80% | 70% | 73% | 83% | 78% | 700 | 11% | ||
dog_bark | 85% | 77% | 81% | 84% | 79% | 81% | 700 | 0% | ||
siren | 82% | 71% | 76% | 85% | 77% | 81% | 602 | 7% | ||
LR | car_horn | 88% | 76% | 82% | 89% | 76% | 82% | 252 | 0% | 16.61% |
children_playing | 54% | 79% | 64% | 72% | 85% | 78% | 700 | 22% | ||
dog_bark | 54% | 81% | 65% | 85% | 80% | 82% | 700 | 26% | ||
siren | 81% | 78% | 79% | 91% | 77% | 84% | 602 | 6% | ||
ANN | car_horn | 72% | 92% | 81% | 84% | 94% | 89% | 252 | 10% | 15.55% |
children_playing | 66% | 71% | 68% | 75% | 82% | 79% | 700 | 16% | ||
dog_bark | 52% | 81% | 63% | 70% | 83% | 76% | 700 | 21% | ||
siren | 69% | 74% | 71% | 89% | 71% | 79% | 602 | 11% | ||
CNN 1D | car_horn | 68% | 85% | 76% | 87% | 90% | 88% | 252 | 16% | 12.31% |
children_playing | 61% | 71% | 65% | 71% | 73% | 72% | 700 | 11% | ||
dog_bark | 56% | 84% | 67% | 73% | 81% | 77% | 700 | 15% | ||
siren | 76% | 70% | 73% | 87% | 74% | 80% | 602 | 10% | ||
CNN 2D | car_horn | 84% | 75% | 79% | 80% | 82% | 83% | 252 | 5% | 8.86% |
children_playing | 64% | 84% | 73% | 72% | 87% | 81% | 700 | 11% | ||
dog_bark | 72% | 79% | 76% | 91% | 76% | 83% | 700 | 9% | ||
siren | 87% | 73% | 79% | 87% | 84% | 85% | 602 | 8% | ||
